Industrial robot health management system and method combining time domain and frequency domain analysis

Abstract

The invention relates to the field of industrial robots, in particular to an industrial robot health management system and method combining time domain and frequency domain analysis. According to the technical scheme, a dynamic time warping algorithm is used for aligning reference current data and real-time feedback current data, then similarity measurement is carried out in a mode of calculating Euclidean distance, whether the industrial robot is in a normal operation state is judged, and after it is judged that the industrial robot has a fault, the fault of the industrial robot is detected. High-pass filtering is carried out on instruction rotating speed data and actual response rotating speed data, and specific faults of the industrial robot are judged in a spectrogram mode; the method has the beneficial effects that whether the operation state of the industrial robot is normal or not can be accurately judged through data alignment; through high-pass filtering and spectrogram modes, the fault type of the industrial robot is accurately judged, and the health management efficiency of the industrial robot is improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial robots, and in particular to an industrial robot health management system and method combining time domain and frequency domain analysis. Background Art

[0002] In the health management and diagnosis of industrial robots, similarity measurement is usually calculated using Euclidean distance. The use of Euclidean distance has a prerequisite, that is, the two time series are completely aligned on the time axis. However, in actual applications, the operating data of industrial robots may shift on the time axis due to factors such as differences in sensor sampling rates, mechanical delays, and external interference. This inconsistency on the time axis will cause the results calculated by Euclidean distance to fail to accurately reflect the actual similarity between signals, resulting in a large deviation.

[0003] In addition, the operation data of industrial robots is affected by a variety of working condition information, such as load changes, operation modes, etc. These factors will cause the spectral characteristics of the signal to change. If the spectrum analysis is directly performed on the operation data affected by the working condition information, a large amount of noise and interference may be introduced, resulting in inaccurate extraction of spectral features. Summary of the invention

[0004] The purpose of the present invention is to provide an industrial robot health management system and method that combines time domain and frequency domain analysis to solve the problems of data misalignment and inability to perform frequency domain analysis on operating data in the current diagnosis process of industrial robots.

[0005] To achieve the above object, the present invention adopts the following technical solution: a method for health management of industrial robots combining time domain and frequency domain analysis, comprising the following steps: S01. First, feedback current of the industrial robot under various normal operating conditions is collected to obtain benchmark feedback current data, and a benchmark data set is constructed.

[0006] S02. Collect the feedback current of the actually running industrial robot in real time to obtain a real-time feedback current data sequence within an operation cycle.

[0007] S03. Retrieving the reference feedback current data of the same operating state as the industrial robot actually operated in step S02 from the reference data set in step S01, and aligning the data with the real-time feedback current data sequence in step S02.

[0008] S04. Calculate the Euclidean distance between the aligned reference feedback current data and the real-time feedback current data in step S03, and determine whether the calculated Euclidean distance exceeds a set threshold. If so, it is considered that the actually running industrial robot has a fault, and proceed to step S05. If not, it is considered that the actually running industrial robot is in normal operation.

[0009] S05. For the situation that is considered to have a fault in step S04, the command speed data and the actual response speed data of the industrial robot in the time period of the real-time feedback current data are obtained, and the command speed data and the actual response speed data are respectively converted from time domain signals to frequency domain signals to obtain frequency spectrum graphs of the command speed data and the actual response speed data; finally, the frequency spectrum graph of the actual response speed data is compared with the frequency spectrum graph of the command speed data to determine the type of fault existing in the actually running industrial robot.

[0010] Specifically, in step S05, the spectrum of the actual response speed data is compared with the spectrum of the command speed data to determine the type of fault of the actually running industrial robot, and the following determination method is specifically used: S51. If there is a vibration peak in the high frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a bearing fault.

[0011] S52. If there are vibration peaks in both the main frequency part and the first harmonic part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a misalignment fault.

[0012] S53. If there are vibration peaks in both the main frequency part and the higher harmonic part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a dynamic balance fault.

[0013] S54. If there are vibration peaks in the gear meshing frequency and its harmonic part and sideband frequency part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a bevel gear fault.

[0014] S55. If there are vibration peaks in both the low-frequency part and the sideband frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a loose installation fault.

[0015] Specifically, in step S05, before converting the command speed data and the actual response speed data from time domain signals into frequency domain signals, the command speed data and the actual response speed data are firstly subjected to high-pass filtering to achieve the purpose of removing the operating condition information.

[0016] Specifically, in step S03, when aligning the reference feedback current data with the real-time feedback current data, a dynamic time warping algorithm is specifically used.

[0017] An industrial robot health management system combining time domain and frequency domain analysis comprises a reference database, a feedback current acquisition module, a data alignment module, a Euclidean distance calculation module, a rotational speed data acquisition module, a spectrum diagram display module and a fault type output module, wherein the reference database is used to store the reference data set acquired by the above method, the feedback circuit acquisition module is used to execute step S02 of the above method, the data alignment module is used to execute step S03 of the above method, the Euclidean distance calculation module is used to execute step S04 of the above method, the rotational speed data acquisition module is used to execute step S05 of the above method to acquire the command rotational speed data and the actual response rotational speed data of the industrial robot within the time period where the real-time feedback current data is located, the spectrum diagram display module is used to display the spectrum diagram of the command rotational speed data and the actual response rotational speed data obtained in step S05 of the above method, and the fault type output module is used to output the fault type existing in the actually running industrial robot obtained in step S05 of the above method.

[0018] The beneficial effects of the present invention are as follows: by using a dynamic time warping algorithm to align the reference current data with the real-time feedback current data, and then calculating the Euclidean distance to measure the similarity, it is possible to accurately judge whether the operating status of the industrial robot is normal; and after judging that the industrial robot has a fault, by high-pass filtering the command speed data and the actual response speed data, and judging the specific fault of the industrial robot by means of a spectrum diagram, the health management efficiency of the industrial robot and the accuracy of fault judgment are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Attached Figure 1 Schematic diagram of system connection of an industrial robot health management system combining time domain and frequency domain analysis in an embodiment. DETAILED DESCRIPTION

[0020] Embodiment 1, a method for health management of an industrial robot combining time domain and frequency domain analysis, comprises the following steps: S01. First, feedback current of the industrial robot under various normal operating conditions is collected to obtain benchmark feedback current data, and a benchmark data set is constructed.

[0021] S02. Collect the feedback current of the actually running industrial robot in real time to obtain a real-time feedback current data sequence within an operation cycle.

[0022] S03. Retrieving the reference feedback current data of the same operating state as the industrial robot actually operated in step S02 from the reference data set in step S01, and aligning the data with the real-time feedback current data sequence in step S02.

[0023] S04. Calculate the Euclidean distance between the aligned reference feedback current data and the real-time feedback current data in step S03, and determine whether the calculated Euclidean distance exceeds a set threshold. If so, it is considered that the actually running industrial robot has a fault, and proceed to step S05. If not, it is considered that the actually running industrial robot is in normal operation.

[0024] S05. For the situation that is considered to have a fault in step S04, the command speed data and the actual response speed data of the industrial robot in the time period of the real-time feedback current data are obtained, and the command speed data and the actual response speed data are respectively converted from time domain signals to frequency domain signals to obtain frequency spectrum graphs of the command speed data and the actual response speed data; finally, the frequency spectrum graph of the actual response speed data is compared with the frequency spectrum graph of the command speed data to determine the type of fault existing in the actually running industrial robot.

[0025] Specifically, in step S05, the spectrum of the actual response speed data is compared with the spectrum of the command speed data to determine the type of fault of the actually running industrial robot, and the following determination method is specifically used: S51. If there is a vibration peak in the high frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a bearing fault.

[0026] S52. If there are vibration peaks in both the main frequency part and the first harmonic part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a misalignment fault.

[0027] S53. If there are vibration peaks in both the main frequency part and the higher harmonic part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a dynamic balance fault.

[0028] S54. If there are vibration peaks in the gear meshing frequency and its harmonic part and sideband frequency part of the spectrum diagram of the actual response speed data compared with the spectrum diagram of the command speed data, it is determined that the industrial robot has a bevel gear fault.

[0029] S55. If there are vibration peaks in both the low-frequency part and the sideband frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a loose installation fault.

[0030] Specifically, in step S05, before converting the command speed data and the actual response speed data from time domain signals into frequency domain signals, the command speed data and the actual response speed data are firstly subjected to high-pass filtering to achieve the purpose of removing the operating condition information.

[0031] Specifically, in step S03, when aligning the reference feedback current data with the real-time feedback current data, a dynamic time warping algorithm is specifically used.

[0032] At the same time, in view of the above method, this embodiment also provides an industrial robot health management system that combines time domain and frequency domain analysis. Figure 1 , which includes a benchmark database, a feedback current acquisition module, a data alignment module, a Euclidean distance calculation module, a speed data acquisition module, a spectrum diagram display module and a fault type output module, the benchmark database is used to store the benchmark data set collected by the above method, the feedback circuit acquisition module is used to execute step S02 of the above method, the data alignment module is used to execute step S03 of the above method, the Euclidean distance calculation module is used to execute step S04 of the above method, the speed data acquisition module is used to execute step S05 of the above method to obtain the command speed data and actual response speed data of the industrial robot within the time period where the real-time feedback current data is located, the spectrum diagram display module is used to display the spectrum diagram of the command speed data and the actual response speed data obtained in step S05 of the above method, and the fault type output module is used to output the fault type of the actually running industrial robot obtained in step S05 of the above method.

[0033] Of course, the above are only preferred embodiments of the present invention, and are not intended to limit the scope of use of the present invention. Therefore, any equivalent changes made to the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for industrial robot health management combining time domain and frequency domain analysis, characterized in that: The steps include: S01. First, collect feedback current of the industrial robot under various normal operating conditions to obtain benchmark feedback current data and construct a benchmark data set; S02, collecting the feedback current of the actually running industrial robot in real time to obtain real-time feedback current data within an operation cycle; S03, calling out the benchmark feedback current data of the same operating state as the industrial robot actually running in step S02 from the benchmark data set in step S01, and aligning the data with the real-time feedback current data in step S02; S04, calculating the Euclidean distance between the aligned reference feedback current data and the real-time feedback current data in step S03, and judging whether the calculated Euclidean distance exceeds a set threshold value, if it exceeds the set threshold value, it is considered that the actually running industrial robot has a fault, and the process proceeds to step S05, if it does not exceed the set threshold value, it is considered that the actually running industrial robot is in a normal operating state; S05. For the situation that is considered to have a fault in step S04, the command speed data and the actual response speed data of the industrial robot in the time period of the real-time feedback current data are obtained, and the command speed data and the actual response speed data are respectively converted from time domain signals to frequency domain signals to obtain frequency spectrum graphs of the command speed data and the actual response speed data; finally, the frequency spectrum graph of the actual response speed data is compared with the frequency spectrum graph of the command speed data to determine the type of fault existing in the actually running industrial robot.

2. The method for industrial robot health management combining time domain and frequency domain analysis according to claim 1, characterized in that: In step S05, the frequency spectrum of the actual response speed data is compared with the frequency spectrum of the command speed data to determine the type of fault of the actually running industrial robot, and the following determination method is specifically used: S51. If there is a vibration peak in the high frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a bearing fault; S52, if there are vibration peaks in both the main frequency part and the first harmonic part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a misalignment fault; S53. If there are vibration peaks in both the main frequency part and the higher harmonic part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a dynamic balance fault; S54. If the gear meshing frequency and its harmonic part and the sideband frequency part of the spectrum of the actual response speed data have vibration peaks compared with the spectrum of the command speed data, it is determined that the industrial robot has a bevel gear fault; S55. If there are vibration peaks in both the low-frequency part and the sideband frequency part of the spectrum of the actual response speed data compared with the spectrum of the command speed data, it is determined that the industrial robot has a loose installation fault.

3. The method for health management of industrial robots combining time domain and frequency domain analysis according to claim 1, characterized in that: In the step S05, before converting the command speed data and the actual response speed data from time domain signals into frequency domain signals, the command speed data and the actual response speed data are firstly subjected to high-pass filtering to achieve the purpose of removing the operating condition information.

4. The method for health management of industrial robots combining time domain and frequency domain analysis according to claim 1, characterized in that: In the step S03, when aligning the reference feedback current data with the real-time feedback current data, a dynamic time warping algorithm is specifically used.

5. An industrial robot health management system combining time domain and frequency domain analysis, characterized in that: The invention comprises a reference database, a feedback current acquisition module, a data alignment module, a Euclidean distance calculation module, a rotation speed data acquisition module, a spectrum diagram display module and a fault type output module, wherein the reference database is used to store the reference data set acquired in step S01 described in any one of claims 1 to 4, the feedback circuit acquisition module is used to execute step S02 described in any one of claims 1 to 4, the data alignment module is used to execute step S03 described in any one of claims 1 to 4, the Euclidean distance calculation module is used to execute step S04 described in any one of claims 1 to 4, the rotation speed data acquisition module is used to execute step S05 described in any one of claims 1 to 4 to acquire the command rotation speed data and the actual response rotation speed data of the industrial robot in the time period where the real-time feedback current data is located, the spectrum diagram display module is used to display the spectrum diagram of the command rotation speed data and the actual response rotation speed data obtained in step S05 described in any one of claims 1 to 4, and the fault type output module is used to output the fault type existing in the actually running industrial robot obtained in step S05 described in any one of claims 1 to 4.

Images